Radar scientists and engineers need to examine and efficiently process large quantities of radar data to verify the performance of new algorithms or to perform quality control on the data. Interactive access to radar data increases productivity by allowing the user to quickly examine various parts of the dataset. In addition, users need to develop their own processing algorithms and easily modify the parameters for processing algorithms. NCAR has developed a Radar Processing Toolkit using Numeric Python to meet these needs. Python, with its Numeric extension, is a freely available, interactive, interpreted language that does not require the compile and link steps needed by compiled languages. Numeric Python processes data efficiently and can easily be extended with high performance subroutines written in C/C++ or FORTRAN. NCAR has used Python's advanced object oriented features to build class libraries that allow scientific users to more easily write their own algorithms.